Since October we have observed a gradual worsening of the PMC transmission when the ITF is in LN3 (look at fig. 1). The comparison between INJ in standalone and LN3 is shown in fig. 4,. so that it seems to be some frequency noise injected by the SSFS (that shows in fact coherence between 100 and 200 Hz, as shown in fig 5)

As a first attempt to come back to the conditions we used to have in October, this morning we adjusted the offset on the locking point of the PMC by tuning the modulation frequency. 

We proceeded as follows: while scanning the PMC we looked at the err+perturb signal (demodulated error signal) on the oscilloscope. The PDH signal was found to have an offset of 500 mV. By changing the modulation frequency from 14.53 MHz to 14.56 MHz the offset went to zero. We then locked again the PMC and we checked we still had  the maximum of transmission (look plot 2). To be noticed that since this frequency had already been tuned from 14.5 to 14.53 MHz in order to reduce the offset (entry 42881) , we should keep monitoring it and try to understand whether it is a systematic drift .

However, the reduction of the offset didn't have any benefits on the trasmission of the PMC (plot 3). This means that the the worsening was not due to frequency to intensity noise conversion and we that need to look for other possible mechanisms. The improvement on the error signal PSL PMC REFL I is probably not linked to this action, but it is due to the fact that we were monitoring the signal with a cable which was connected to the DAQ, and (by mistake) also in parallel to a buffer.